#' @title get Forecast
#'
#' @description Predict time-series curves for the selected days indices (lines in data).
-#' Run the forecasting task described by \code{delta_forecaster_name} and
-#' \code{shape_forecaster_name} on data obtained with \code{getData}
#'
#' @param data Dataset, object of type \code{Data} output of \code{getData}
#' @param indices Days indices where to forecast (the day after)
-#' @param memory Data depth (in days) to be used for prediction
-#' @param horizon Number of time steps to predict
-#' @param shape_forecaster_name Name of the shape forcaster
+#' @param forecaster Name of the main forcaster
#' \itemize{
#' \item Persistence : use values of last (similar, next) day
-#' \item Neighbors : use PM10 from the k closest neighbors' tomorrows
+#' \item Neighbors : use values from the k closest neighbors' tomorrows
#' \item Average : global average of all the (similar) "tomorrow of past"
+#' \item Zero : just output 0 (benchmarking purpose)
+#' \item Level : output a flat serie repeating the last observed level
#' }
-#' @param delta_forecaster_name Name of the delta forecaster
+#' @param pjump How to predict the jump at the interface between two days ?
#' \itemize{
#' \item Persistence : use last (similar) day values
-#' \item Neighbors: re-use the weights optimized in corresponding shape forecaster
+#' \item Neighbors: re-use the weights optimized in corresponding forecaster
#' \item Zero: just output 0 (no adjustment)
#' }
+#' @param memory Data depth (in days) to be used for prediction
+#' @param horizon Number of time steps to predict
#' @param ... Additional parameters for the forecasting models
#'
#' @return An object of class Forecast
#'
#' @examples
#' data = getData(ts_data="data/pm10_mesures_H_loc.csv", exo_data="data/meteo_extra_noNAs.csv",
-#' input_tz = "Europe/Paris", working_tz="Europe/Paris", predict_at="07")
-#' pred = getForecast(data, 2200:2230, Inf, 12, "Persistence", "Persistence")
+#' input_tz = "Europe/Paris", working_tz="Europe/Paris", predict_at=7)
+#' pred = getForecast(data, 2200:2230, "Persistence", "Persistence", 500, 12)
#' \dontrun{#Sketch for real-time mode:
#' data = new("Data", ...)
+#' forecaster = new(..., data=data)
#' repeat {
#' data$append(some_new_data)
-#' pred = getForecast(data, ...)
+#' pred = forecaster$predict(data$getSize(), ...)
#' #do_something_with_pred
#' }}
#' @export
-getForecast = function(data, indices, memory, horizon,
- shape_forecaster_name, delta_forecaster_name, ...)
+getForecast = function(data, indices, forecaster, pjump,
+ memory=Inf, horizon=data$getStdHorizon(), ...)
{
+ # (basic) Arguments sanity checks
horizon = as.integer(horizon)[1]
if (horizon<=0 || horizon>length(data$getCenteredSerie(2)))
stop("Horizon too short or too long")
if (any(indices<=0 | indices>data$getSize()))
stop("Indices out of range")
indices = sapply(indices, dateIndexToInteger, data)
-
- #NOTE: some assymetry here...
- shape_forecaster = new(paste(shape_forecaster_name,"ShapeForecaster",sep=""), data=data)
- #A little bit strange, but match.fun() and get() fail
- delta_forecaster = getFromNamespace(
- paste("get",delta_forecaster_name,"DeltaForecast",sep=""), "talweg")
+ if (!is.character(forecaster) || !is.character(pjump))
+ stop("forecaster and pjump should be of class character")
pred = list()
+ forecaster = new(paste(forecaster,"Forecaster",sep=""), data=data,
+ pjump = getFromNamespace(paste("get",pjump,"JumpPredict",sep=""), "talweg"))
for (today in indices)
{
- #shape always predicted first (on centered series, no scaling taken into account),
- #with side-effect: optimize some parameters (h, weights, ...)
- predicted_shape = shape_forecaster$predict(today, memory, horizon, ...)
- #then, delta prediction can re-use some variables optimized previously (like neighbors infos)
- predicted_delta = delta_forecaster(data, today, memory, horizon,
- shape_forecaster$getParameters(), ...)
-
- #TODO: this way is faster than a call to append(); why ?
pred[[length(pred)+1]] = list(
- # Predict shape and align it on end of current day
- serie = predicted_shape + tail( data$getSerie(today), 1 ) - predicted_shape[1] +
- predicted_delta, #add predicted jump
- params = shape_forecaster$getParameters(),
- index = today )
+ "serie" = forecaster$predict(today, memory, horizon, ...),
+ "params" = forecaster$getParameters(),
+ "index" = today
+ )
}
new("Forecast",pred=pred)
}